← Back to NIKSHA OS

NIKSHA OS — AI Usage & Disclaimer

Version 2.0 · Last updated 7 October 2026

Operator: NIKSHA Technologies ("NIKSHA OS", "we", "us").

This document explains, in plain language, how NIKSHA OS uses artificial intelligence (AI), what AI does and does not do with your data, and the limits of AI output. It supports the Privacy Policy and the Acceptable Use Policy.

1. Where AI is used

Corrected 8 August 2026. An earlier version of this document described AI as simply "optional". That understated the position: AI features are switched ON by default for each school. A school can have them limited or switched off, but they are not off until someone asks. This matches Privacy Policy §3.6.

NIKSHA OS offers AI-assisted features that help staff and parents work faster. They are enabled by default per school and can be disabled by the Institution.

• The in-app assistant answers most questions itself, from the school's own records, using NIKSHA OS's own software — no AI model is involved and nothing is sent to an AI provider. For parents and students, the assistant never sends your words to an AI model. For school staff, a question it cannot answer that way may be sent, with the school data needed to answer it, to a third-party AI provider (§2).
• The other AI features listed below send the relevant school data — which can include students' names, attendance, marks or fee status — to the AI provider to write their text.
• Your conversation with the assistant is stored so you can see your own history. It is erased when you delete your account.

These currently include:

• An assistant / copilot that answers questions and drafts text within the app.
• Parent insights — plain-language summaries of a child's attendance, results and progress for the parent.
• Question-paper assistance — suggesting questions to fill gaps in a draft paper, drawn from the school's own question bank and syllabus.
• Admissions assistance — summarising the school's own enquiry/lead funnel and suggesting next actions.
• Director / executive summaries — narrative summaries of the school's own operational metrics.
• Predictions — risk/likelihood scores (e.g. fee-default or admission conversion). The score itself is calculated by fixed rules, not by AI; an AI model may write a plain-language explanation of a score that is already set.
• Marketing / poster captions — draft captions and copy for school posts.

Every AI feature is designed with a safe fallback: if the AI service is not available, the feature uses deterministic (non-AI) logic or is hidden, so the app keeps working.

2. Which AI provider we use

• The AI provider is Anthropic (USA, the "Claude" family of models) by default. The platform operator can instead route requests through OpenRouter (USA) or OpenAI (USA); the Sub-processors list is updated before that happens. Requests are sent over an encrypted connection.
• The provider processes the text we send to produce a response. Your conversation is stored by NIKSHA OS so you can see your own history; it is erased when you delete your account.
• We do not authorise the provider to use your data to train its models. How the provider actually handles what it receives is governed by that provider's own terms and our contract with it, which we cannot vary unilaterally. Under Anthropic's published commercial API terms, it may not train models on the content we send and deletes inputs and outputs within 30 days, except where it keeps flagged content longer (up to 2 years) to enforce its usage policy or must keep data by law (see Privacy Policy §3.6).
• Generated answers may be cached for at least 24 hours inside the school's own data so the same answer is not generated twice (see the Retention Policy §2.6).
• The provider's servers may be located outside India. We send only the minimum text needed for the feature, drawn from data the requesting user is already authorised to see.
• If the platform has no AI provider configured, or the school has switched AI off, no data is sent to any AI provider and features fall back to non-AI logic.

See the Sub-processors list for the current AI provider and its role.

3. What AI does NOT do

• ❌ It does not see data the requesting user isn't already allowed to see — AI runs within the same role-based access controls as the rest of the app.
• ❌ It does not build advertising or marketing profiles of children or families.
• ❌ It does not make automated decisions about a person. AI output is a draft or suggestion for a human to review and act on. (Automatic reminders — for example a fee-due or absence alert — are triggered by plain facts such as a due date or an attendance mark, not by AI.)
• ❌ It is not used for any tracking, behavioural monitoring or targeted advertising directed at children.

4. Important disclaimer — AI can be wrong

AI features are assistive tools, not authoritative sources. AI-generated text, summaries, scores, predictions and suggestions:

• may be inaccurate, incomplete or out of date, and
• must be reviewed by a responsible human before being relied on, shared with parents, published, or used to make any decision about a student, parent or staff member.

Do not treat AI output as a final exam mark, an official record, financial advice, legal advice, medical advice, or a guaranteed prediction. Final responsibility for any academic, financial, disciplinary or administrative decision always rests with the school and the authorised human user — never with the AI.

Predictions and risk scores are statistical estimates to help prioritise attention; they are not judgements about a person and must not be used to unfairly disadvantage any student.

5. Human oversight & content governance

• AI suggestions for sensitive workflows (for example, question papers and published posters) pass through an approval/moderation step before they go live, and rejected or pending items are not published or exported.
• Schools remain responsible for what they approve and publish using AI assistance, and must comply with the Acceptable Use Policy.

6. Your choices

• AI features are on by default for each school, and a school can ask for them to be limited or switched off.
• If you have concerns about an AI-generated output, raise them with your school and, if needed, contact support@nikshaos.in.

7. Changes

As AI features and the law evolve (including any future AI-specific regulation), we will update this disclaimer. Material changes are recorded in the CHANGELOG.